Skip to contents

Shape Distance Between Two Curves

Source: R/alignment.R
shape.distance.Rd

Compute the elastic shape distance between two curves in a specified quotient space. The shape distance factors out the specified nuisance transformations (reparameterization, translation, scale).

Usage

shape.distance(
  f1,
  f2,
  argvals = NULL,
  quotient = c("reparameterization", "translation", "scale"),
  lambda = 0
)

Arguments

f1

Numeric vector (first curve).

f2

Numeric vector (second curve).

argvals

Numeric vector of evaluation points. If NULL, defaults to seq(0, 1, length.out = length(f1)).

quotient

Character: the quotient space. One of "reparameterization", "translation", or "scale".

lambda

Regularization parameter (default 0).

Value

A list with components:

distance

Shape distance

gamma

Optimal warping function

f2.aligned

Aligned version of f2

References

Srivastava, A. and Klassen, E. (2016). Functional and Shape Data Analysis. Springer.

See also

shape.representative, shape.mean, shape.distance.matrix

Examples

# \donttest{
t <- seq(0, 1, length.out = 50)
f1 <- sin(2 * pi * t)
f2 <- 2 * sin(2 * pi * t^1.5) + 1
d <- shape.distance(f1, f2, argvals = t, quotient = "scale")
d$distance
#> [1] 0.2732183
# }